powerdns: Update to version 5.1.3
5.1.3
Released: 30th of June 2026
This is release 5.1.3 of the Authoritative Server.
It contains one new double feature and a few bug fixes.
New Features
Implement SOA-EDIT spreading + RRSIG expiry extension setting.
Bug Fixes
REST API: escaping issue with generic record syntax
pdnsutil: non-interactive zone edit
set version in meson.build in autoconf make dist path
declare enable-lua-record-updates unconditionally
meson: fix missing symbols for backend dylibs
wadptr: Update to 3.8
- Remove existing pkgsrc patch (merged upstream)
- Add pkgsrc patch for tolower() parameter of type char
- Add pkgsrc patch to install files with desired permissions
- Force C standard to C99
- Enable testsuite
- Take maintainership
* The `-u` ("uncompress") option was renamed to `-d` ("decompress", the
usual terminology used by almost every other compression tool).
* A bug was fixed that made the Windows version for the previous release not
work properly; a dependency was unintentionally added on the pthreads API,
the DLL for which was not included in the release. That dependency has now
been removed. Thanks to Anonysussy and happygreenfairy for the bug report.
* A special sector tag range (9700-9799) is now recognized for special
effects. There are some cases where level authors may deliberately want
sidedefs to be merged even when they are attached to special lines, and
this allows that. Thanks to kitchen-ace for the suggestion.
[16 lines not shown]
py-asyncssh: updated to 2.24.0
2.24.0 (27 Jun 2026)
* Added support for creating and validating SSHSIG signatures, as well
as OpenSSH "allowed signers" files.
* Added support for ML-KEM key exchange from the PyCA cryptography
package when it is available. This avoids the need to have the
liboqs library installed to use ML-KEM. However, liboqs is still
required to support SNTRUP kex exchange.
* Fixed an issue with handling of SSH maximum packet size when opening
a new SSH channel, aborting with a protocol error if a peer attempts
to set this size to 0. Thanks go to GitHub user afldl for reporting
this issue and providing analysis and reproduction code.
* Fixed an issue with include directives in OpenSSH config files.
Thanks go to GitHub users sethholmes and tazle for reporting this
issue and providing analysis and a proposed fix.
/bin/sh fix trap/jobid builtins when run from grandchild
The trap and jobid builtins are designed to allow
T=$(trap)
(etc) to see (output) the traps (or jobs for jobid) of parent shell (the one
running the assignment) to be seen from inside a subshell (here the command
substitution, which it usually would be).
But all that should be reported are the traps (jobs) of the parent of
that subshell, not recursively up the chain.
That is
( T=$(trap) )
should usually be equivalent to T='' as inside a subshell environment
all non-ignored traps are defined to be reset to their defaults
[33 lines not shown]
mk/tools/defaults.mk: Document that TOOL bash must be >= 4.0
macOS ships bash 3 (despite bash4 being released in 2009), and pkgsrc
had a few accomodations for packages that needed bash where bash 3 did
not work. pkgsrc no longer uses macOS system bash as a tool, as such
workarounds being in packages is not reasonable.
Requiring system bash >= 5 to allow use as TOOL would seem reasonable,
as bash 5.0 was released in 2019. However, SmartOS uses bash 4, and
years of bulk builds have turned up no problems from bash 4 as a tool.
Therefore, there is no reason to exclude 4, as the point of setting
minimums is to limit human effort and secondarily reduce needless
machine resource use.
While this commit only adds a comment, it provides a basis to
- remove any workarounds for bash 3 as a tool
- disable TOOLS_PLATFORM.bash settings on platforms where bash is < 4
- add a workaround to a package if it needs TOOL bash but is not ok
with bash 4
[2 lines not shown]
sh tests: builtins: add more content
Many of the t_builtins sh tests were just placeholders for things to come.
They've been that way for years...
This adds some actual content to some of them - still very primitive in
most cases, but better than was there before. There are still too many
empty tests (test the builtin exists, but that's it).
Also, add a test for the trap builtin - that wasn't there before as traps
are tested in other of the sh tests ... this new one is therefor not testing
that traps work (or not in any way that's meaningful), rather it is testing
that the various ways the trap command can be used all work.
Currently two of the sub-tests will fail, the new trap test, and the
(was previously empty) jobid test (testing a NetBSD extension builtin).
Adding those failing tests was the inspiration for all of this. The bug
they expose is exotic, and never likely to be encountered, or do any harm
[7 lines not shown]
wadgadget: Update to 0.2
New features:
- Mouse input is now suppported, and many parts of the UI can now be
interacted with by clicking.
- A texture editor was added, making the TEXTURE browser
feature-complete.
- A browser/editor for GENMIDI lumps was implemented, including export
to / import from .sbi voice files.
- A new action was added to update existing WAD entries. This works
the same as the import action but instead automatically replaces the
contents of existing entries, rather creating new ones.
- A new action to move files between directories was added.
- Duke3D/ROTT .rts files are now recognized as WAD files, and embedded
.voc files (found inside them) are recognized as sound effects.
- It is now possible to edit the headers of graphic lumps (useful in
particular in order to change the X/Y offsets).
- It's now possible to import graphics from a variety of other file
formats in addition to .png: .pcx, .lbm, .gif, .bmp, .ppm, .tiff and
[27 lines not shown]
sh tests: arithmetic: fix test for support of pre ++/-- ops
The C prefix and postfix unary ++ and -- operators are optional
in POSIX sh (supported in NetBSD's sh). To allow these tests to be
used to test other shells, the tests generally attempt to verify that
optional features are supported, before testing them.
The prefix ++ and -- tests, to see if there was support in ${TEST_SH}
were not guaranteed to work, as ++X might be preincrement X, or it
might just be + + X (ie: applying the unary + operator twice).
Fix that by testing whether ++X and + + X give the same result, if
they do, then prefix ++ is not supported (also if ++X generates an error)
Just the same for the prefix -- operator.
The postfix operator tests are OK, X-- is either X-- or an error.
NFCI for tests of the NetBSD sh. (This change has been in my local
set of sh tests for a long time now.)
